CDOM/CM Reset

[wpetilli]

All my remote sites are LSP's off a core running 6.0.1. When I log into the system platform of the local s8300 I see the uptime of cdom and cm as time frames that don't make much sense. A few sites say 5days, others longer/shorter. There has been no power issues or anything related to those sites for that to not have a very long uptime. Does that do some sort of maintenance while it's dormant or is it something I should have checked out?

 

[Hudson22]

Are you sure that you are not looking at CPU time instead? CPU Time= The amount of CPU time the VM had since boot.
Hudson22

 

[wpetilli]

On system platform virtual machine management (machine list) I am looking at all the components (domain-0, cm, bsm, cdom, utility server) and in the cpu time column.. for instance cm says 7h 17m 56s. Does that mean cm rebooted 7hours ago, because I've had no reboots or maintenance of any kind. Utility Server says 1h 6m 52s.. cdom 5h 21m 37s..

 

[montyzummer]

This is mis-leading here is a description of the CPU TIME from the PDF

The amount of CPU time the virtual machine has had since boot. This is
not the same as uptime.

So in answer to your question no nothing has rebooted , It looks like the amount of time that the CPU has had to do something out of normal operation (when you are configuring , etc ????) I am not really sure maybe somebody else can clarify the exact meaning , my cm virtual machine has been running for 6 months with no reboot but platform says 9days 18hours.

If you want to be able to prove this to a customer you could pull the logs under server management log viewer and time stamp it with the days counted back from the CPU time.


APSS (SME)
ACSS (SME)
ACIS (UC)

 

[wpetilli]

Ok, so I have to bring this one back up. I had a scheduled power down in a remote office this weekend. The site is an s8300/G450 (cm6.0.1). A half hour before the equipment was being turned off I logged into System Platform and wanted to gracefully shut down the s8300 before they pulled the plugs on the GW. I logged in and went to 'Server Management' and did a 'shutdown', not 'reboot' of the system. I 1000% know I did shutdown. I confirmed and had a continuous ping up and watched the replies time out. I logged into my main CM to make sure I saw it offline using 'list surv' command. 15 minutes later, the pings started replying and the 8300 was back online. I ran out of time and didn't want to muck with it so I let the onsite guy pull the plugs. It came up fine the next day, but I do have to ask if there's a specific way to shut this thing down that I'm not doing?

 

[montyzummer]

did the guy on site ensure the ok to remove light was steady ??on the s8300

APSS (SME)
ACSS (SME)
ACIS (UC)

 

[wpetilli]

I didn't have anyone bother to even look at it. I just figured a shutdown server command and watch the pings time out were enough.

 

[montyzummer]

Well I would say , that is a definitive answer that's what its there for I guess , I have had the system stop responding to ping when the light is not steady , also if you want to check cm ,putty into cm ip and type server status.

APSS (SME)
ACSS (SME)
ACIS (UC)

 

[wpetilli]

This is the 2nd site the virtual machines all came back online after a shutdown. I'm wondering if on the main dashboard of sys plat, if you click on dom0 and do the shutdown from within there. In the pre system platform days, doing shutdown from the cm web interface always did the trick.

 

[montyzummer]

to be sure of the state of dom O go into cli su to root type xm list and it will say r if its running , tbh I don't know the physical command to shutdown dom 0 from cli but im sure somebody else will , interesting that, it has restarted the whole lot from a shutdown command from the web interface as to me shut down is what it says on the tin , but as I said the light is the key on the embedded boys.



Hang on of course if these guys have pulled the plugs the whole environment will restart when power is re applied , I guess you just want to check that everything is down before the plug is pulled correct??

APSS (SME)
ACSS (SME)
ACIS (UC)

 

[wpetilli]

I just wanted to gracefully shutdown the 8300 before the plugs were pulled on the GW. We have so many remote locations and there is some sort of power maintenance going on all the time. Since the GW is registered to the core, I sometimes like to just shutdown the 8300 way earlier than the GW plugs being pulled since there would be no impact and I don't have to be at a console at some weird time of the weekend/night.

 

[montyzummer]

you could putty into cm type stop -a this would stop all processes then keep checking server status to be really sure .But the web interface of cm should and can be trusted.

APSS (SME)
ACSS (SME)
ACIS (UC)

 

[wpetilli]

but is going to the web interface of CM enough to shutdown the server or does it have to be done from Sys Plat?

 

[montyzummer]

that is only going to shut down cm the commands I have given you , to shut down the actual hypervisor(xen ) dom0 and cdom its the little button on the server or indeed the method you mentioned before from cdom web interface , im sure in root of dom 0 you can shut down every vm separately using the xm ??? commands but google and a test bench is the best place for that.

In my mind if I shut down a s8300d its the button and steady light , if using the web interface check the light if its steady its worked simple.

APSS (SME)
ACSS (SME)
ACIS (UC)